A strain gauge measures strain in a material to which it is attached and can also be used to build load cells and accelerometers. Early strain gauges were manufactured by winding a thin wire around a paper tube and bonding the wire to the tube with cellulose glue. This wire wound combination was then flattened and bonded to the material to be tested for strain. The wire wound strain gauge was thick and limited by the type of adhesives available. Glue lines caused inaccuracies in strain measurements and caused creep when the gauge was used in load cells. Because the wire was wrapped in a spiral pattern and then flattened, a small length of the wire would be perpendicular to the length of the gauge at each end of the gauge. This small length of wire was sensitive to transverse strain, which is strain perpendicular to the strain direction that was intended to be measured by the gauge. The sensitivity to transverse strain is an undesirable feature for a strain gauge used in an unknown strain field because the angle of the desired strain to be measure with respect to the position of the gauge is unknown. The output signal generated from the transverse strain, therefore, cannot be separated from the output signal generated from the desired strain. This results in an error in the strain measurement values.
Following the wire wound strain gauge was the development of etched metal foil strain gauges. The etched foil gauge was manufactured from a metal foil about 5 .mu.m thick bonded to a special plastic backing about 50 .mu.m thick. A pattern was then etched into the foil to form a grid which was sensitive to strain in one direction. The etched foil gauges were found to be more sensitive to transverse strain than the wire wound gauges. This occurs because the relative width of the etched grids compared to the foil thickness allows inducement of strain in the grid across its width, which in turn changes the known resistance value of the gauge. Therefore, an error in measurement of strain occurs because the resistance change of the gauge due to the transverse strain is unknown. The etched foil gauges also have small transverse sections at each end of the grid that respond to transverse strain causing further errors in strain measurement.
Neither the wire wound gauges nor the etched foil gauges can be calibrated for use. For either type of gauge to work properly the part of the gauge bonded to the material to be measured must be strained by the strain field of the material. This requires the gauge to be flexible so when it is bonded to the material being measured it will transmit the strain into the metal foil or the wire. Once the gauge has been bonded to the material to be measured, it cannot be removed without damaging the gauge. This prevents the gauge from being mounted in a known strain field for calibration and then removed so it can be bonded to the material with the unknown strain field. Both the wire wound and the etched foil gauges are calibrated by selecting a small sample of gauges from a batch of manufactured gauges. The samples are then sacrificed by mounting them in a known strain field and measuring their output signal. From this signal an average gauge factor is determined and assumed that it applies to all the other gauges of that batch for calibration purposes.
The above mentioned strain gauges have been used in either a quarter, half or full Wheatstone bridge circuit configuration. A quarter Wheatstone bridge circuit is a full Wheatstone bridge with only one strain gauge, where the full Wheatstone bridge has four strain gauges. For a quarter Wheatstone bridge, the other three legs of a full Wheatstone bridge have precision resistors to replace the three strain gauges usually present. This configuration is required for the measurement of an unknown strain field because there is no way to determine the signal contribution of each gauge if there is more than one strain gauge. Wire wound and etched foil strain gauges also change resistance with temperature. With a quarter Wheatstone bridge, the active gauge will generate a signal that is based on temperature of the material to be measured. The signal of the active gauge due to temperature cannot be separated from the signal due to strain, thereby causing an error in the strain measurement. Since the other legs are not strain gauges, they do not respond to temperature in the same manner as the strain gauge and cannot be used to eliminate the signal due to temperature.
A half Wheatstone bridge circuit has two strain gauges. This circuit is used in some transducers, but cannot be used for the measurement of strain in an unknown strain field because of having more than one strain gauge as stated above. In the full Wheatstone bridge circuit, where all four legs are strain gauges, the signal due to temperature of the material cancels. Also, a larger strain signal is generated by the combination of four strain gauges. Another advantage of the full Wheatstone bridge circuit is that for much larger strains the output signal due to strain is linear, where the signal from quarter and half bridge circuits are not linear for larger strains. All the above mentioned features makes the full Wheatstone bridge a desirable circuit for strain measurement. However, the full Wheatstone bridge circuit cannot be used in an unknown strain field because there is no way to determine the signal contribution of each of the four gauges.
The next development in strain measurement was the semi-conductor strain gauge. Semi-conductor strain gauges do not produce an output signal that is linear with strain of the material to be measured, but by controlling the doping of the silicon crystal, a much larger output signal is produced for a given strain than a wire wound or etched foil gauge due to the material used in manufacture. However, they are much more temperature sensitive than the etched foil or the wire wound gauges. The semi-conductor gauges are usually used for unsteady force measurements where long term drift is not important and very small strains are to be measured. These gauges have the same adhesive limitations as the etched foil and wire wound gauges. There is also a change in the known resistance of the gauge due to the contraction of the adhesives as they cure because the semi-conductor gauges are so sensitive. Another type of strain gauge is made in the form of a transducer, which is manufactured using a process of thin film deposition. This is similar to electroplating except it is done in a vacuum at high temperature. The advantage of thin film deposition is that the film forms a molecular bond with the material it is being deposited onto. A thin film of metal is usually deposited onto a sensing element in a transducer that has been coated with an electrical insulator. The film is then etched into four foil strain gauge patterns to produce a full Wheatstone bridge circuit configuration. This technique eliminates the need for an adhesive bond but the transducer is limited by a smaller output signal from the strain, due to the low response inherent of any gauge using metal foil.
One of the desired features of a strain gauge is a large output signal due to the strain to be measured because the large signal is easier to measure. The output signal of strain from all the above mentioned strain gauges is dependent on the material used to manufacture the gauge and the whether a quarter, half or full Wheatstone bridge circuit configuration is used. As discussed above, the full Wheatstone bridge gives the largest output signal of strain plus the other important advantages inherent to this circuit. So, in order to achieve the largest possible signal and the other advantages involved with the full Wheatstone bridge, a way to used such a bridge circuit is desired.
It is the objective of this invention to provide a strain gauge that has lower inherent errors than previous strain gauges, is insensitive to transverse strain, provides a large output signal and employs the desired attributes of a full Wheatstone bridge circuit.